Aims Programmed death-ligand 1 (PD-L1) is known to be highly expressed in various malignancies, including head and neck squamous cell carcinoma (HNSCC). We aimed to determine the prevalence of PD-L1 expression in recurrent or metastatic HNSCC (R/M HNSCC) among Chinese patients.
Methods This multicentre, retrospective analysis of data from six centres in China included patients with R/M HNSCC treated from 9 August 2021 to 28 February 2022. PD-L1 expression in tumour tissue was assessed and represented using a combined positive score (CPS). The χ2 and Cochran-Mantel-Haenszel χ2 tests were used to compare the prevalence of different PD-L1 expression statuses according to related co-variables.
Results For all 402 examined patients with R/M HNSCC, 168 cases (41.8%) had PD-L1 expression with a CPS ≥20, and 337 cases (83.8%) had PD-L1 expression with a CPS ≥1. Between the PD-L1 CPS ≥20 group and PD-L1 CPS <20 group, statistically significant differences were observed for variables of sex (p<0.001), smoking habit (p=0.0138 for non-smokers vs current smokers) and primary tumour site (p<0.001 for hypopharynx vs oral cavity and p=0.0304 for larynx vs oral cavity, respectively).
Conclusion PD-L1 with CPS ≥20 was expressed in about 41.8% of cases with R/M HNSCC among Chinese patients, and PD-L1 expression was significantly associated with sex, smoking history and primary tumour site. Our findings regarding the variables related to PD-L1 expression level provide insight for clinical practice and a solid basis for future research on immunotherapy in HNSCC.
Trial registration number ISRCTN10570964.
- CELL BIOLOGY
Data availability statement
Data are available upon reasonable request. The data generated in this study are available upon request from the corresponding author.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
WHAT IS ALREADY KNOWN ON THIS TOPIC
Head and neck squamous cell carcinomas (HNSCCs) are reported to be the sixth most common cancer worldwide. Programmed death-ligand 1 (PD-L1) expression is the guiding factor for the use of immune checkpoint inhibitors, but the PD-L1 expression status in Chinese patients with HNSCC is largely unknown.
WHAT THIS STUDY ADDS
This study provided data regarding the expression of PD-L1 among Chinese patients with HNSCC as well as how PD-L1 expression is associated with sex, smoking history and primary tumour site.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
The insights derived from this study’s data may help inform the use of immune checkpoint inhibitors and serve as a basis for future research on immunotherapy in HNSCC.
Head and neck squamous cell carcinoma (HNSCC) is reported to be the sixth most common cancer worldwide.1 Furthermore, according to the National Central Cancer Registry of China, 142 000 new cases of HNSCC occurred in China, along with nearly 75 000 deaths, in 2020.2 3 Around half of cases with HNSCC have regional lymph node involvement at diagnosis (ie, recurrent or metastatic HNSCC (R/M HNSCC)) with poor prognosis: the median survival is only about 6–15 months depending on patient-related and disease-related factors.4
Programmed death ligand 1 (PD-L1) is the principal ligand of programmed death 1 (PD-1).5 Additionally, PD-L1 can prevent the hyperactivation of uncontrolled T cells and protect normal cells.6 PD-L1 can be found in malignant cells, including HNSCC, but was not found to be associated with poor prognosis in previous studies.7–9 However, it was reported that PD-1/PD-L1 checkpoint inhibitors (nivolumab and pembrolizumab) have remarkable antitumour activity against PD-L1-expressing tumours.10 11 Therefore, PD-L1 expression is the guiding factor for the use of immune checkpoint inhibitors.12 Both the prevalence and prognostic value of PD-L1 expression in HNSCC are still subjects of contention, though several studies have reported that patients with PD-L1-positive tumours are more likely to respond to PD-1/PD-L1 inhibitors.13–17 Moreover, the most recent results from the KEYNOTE-048 study showed that pembrolizumab was effective as a first-line treatment for patients with R/M HNSCC with PD-L1-positive cancer.18
The expression of PD-L1 on non-tumour cells in the tumour microenvironment has been described in some studies.19 20 However, the distribution of PD-L1 on non-malignant cells in HNSCC and the clinical phenotype most likely to benefit from immunotherapy remain unclear.21 Additionally, the main cause of HNSCC in Europe and the USA is oral human papillomavirus (HPV) infection, but in China, smoking or drinking are the leading causes.1 It is necessary to collect solid evidence regarding PD-L1 expression in HNSCC for distinguishing the predictive subgroups more likely to benefit from different treatments. A combined positive score (CPS), which was defined in the KEYNOTE-055 study as ‘percentage of tumour and mononuclear inflammatory cells within the tumour nests and adjacent supporting stroma expressing PD-L1 at any intensity’,22 was proposed to be the most appropriate score for determining PD-L1 expression in HNSCC.23 Because the PD-L1 expression status in Chinese patients with HNSCC is largely unknown, our study aimed to determine the prevalence of PD-L1 high expression (determined by CPS and assessed by immunohistochemistry (IHC) with PD-L1 IHC 22C3 pharmDx kit on ASL48 staining platform) in Chinese patients with advanced R/M HNSCC.
This study was a multicentre retrospective analysis of data from six centres in China. The list of participating centres can be found in online supplemental appendix 1. The protocol and protocol amendments of this study were reviewed and approved by each study site’s Independent Ethics Committee. All patients signed the informed consent form (ICF) before enrollment into the study. The present trial was registered at the ISRCTN registry.
We screened patients with R/M HNSCC in six participating centres in China from 9 August 2021 to 28 February 2022. The inclusion criteria included: (1) having a histologically or cytologically confirmed diagnosis of R/M HNSCC that was considered incurable by local therapies (patient could not have a primary tumour site of the nasopharynx with any histology); (2) age ≥18 years at diagnosis of R/M HNSCC; (3) having an available formalin-fixed, paraffin-embedded (FFPE) tumour specimen obtained via core or excisional biopsy; (4) missing data for related medical information less than 5%24–26 and (5) previously providing written informed consent form, giving consent for his/her sample to be used in a future study, unless the patient was under conditions accepted by the institutional review board/ethics review committee(IRB/ERC) to waive the need for an ICF. Otherwise, the patient must have provided a specific written informed consent for this study. The exclusion criterion of the patients was who have only a specimen obtained via fine needle aspirate (FNA) or a cytologic specimen. Notably, AJCC Cancer Staging Manual, 8th edition was used for disease stage.27
The samples for PD-L1 measurement were FFPE tumour specimens obtained via core or excisional biopsy, which represented a newly obtained biopsy specimen (within 90 days prior to start of PD-L1 (IHC) evaluation), or at least the archival tissue block had to be no older than 2 years. Regarding the sample site, we preferred tumour specimens collected from the primary site rather than the metastatic site. Furthermore, decalcified bony specimens were not accepted.
A minimum of three freshly cut (within 1 month of assay, prefer within 1 week) 4 µm sections from selected FFPE specimens were required for each patient. Two unstained slides were used for PD-L1 detection and negative control, and one would be stained with H&E to assess tissue histology and preservation quality. However, four unstained slides and one matched H&E-stained slide were preferred to allow for rapid retesting. The detailed procedures for use of the PD-L1 IHC 22C3 pharmDx kit (Agilent Technologies, Santa Clara, California, USA) for tissue collection, processing and shipment were consistent with those applied in the KEYNOTE-048 study.18 The IHC analysis and CPS scoring were performed locally at each site by a pathologist who had been trained and certified through a training programme. Details regarding PD-L1 expression analysis, including the date of analysis and CPS, were electronically captured in a central database and merged with other key variables.
The PD-L1 expression level was determined in all samples using the PD-L1 IHC 22C3 pharmDx kit. Clone 22C3 is intended for detection of PD-L1 protein in FFPE tissues using the EnVision FLEX visualisation system on Autostainer Link 48. PD-L1 protein expression was represented by the CPS, which is the number of PD-L1-stained cells (tumour cells, lymphocytes, macrophages) divided by the total number of viable tumour cells, multiplied by 100, according to the following equation:
The primary outcome was the prevalence of PD-L1 CPS ≥20 expression in R/M HNSCC. The secondary outcomes were the prevalence of PD-L1 CPS ≥1 expression in R/M HNSCC and the factors associated with expression of PD-L1 in R/M HNSCC.
We collected relevant demographic characteristics (eg, age at diagnosis, sex, family history of the studied disease and history of tobacco use), clinicopathological parameters (eg, primary tumour site, tumour stage, histology and grade, metastatic location and number, site and type of tumour tissue sample), details of treatment status (eg, previous lines of therapy and prior curative treatments) and other available biomarkers (eg, HPV status) from each centre’s electronic medical records (EMR) system or by chart review if no EMR existed.
Tobacco and alcohol consumption
The definitions used to define the alcohol consumption status of patients are as follows:
Alcohol dependent, sometimes known as ‘alcoholism’, is the most serious form of drinking problem and describes a strong, often uncontrollable, desire to drink.
Heavy drinker: a person who has frequent episodes of heavy drinking (>5 drinks at a time), which can be defined by sex and periodicity.
Occasional drinkers: they drank less than 1 day per week (for instance once or twice per month) and drinking a maximum of 3 drinks for men and 2 drinks max for women. Binge drinkers were incorporated into the heavy drinker category.
Abstinent: abstinence from alcohol lasting 6 months or longer.
The definitions used to define the tobacco consumption status of patients are as follows:
Current smoker: an adult who has smoked 100 cigarettes in his or her lifetime and who currently smokes cigarettes.
Former smoker: an adult who has smoked at least 100 cigarettes in his or her lifetime but who had quit smoking at the time of interview.
Non-smoker: an adult who has never smoked, or who has smoked less than 100 cigarettes in his or her lifetime.
To decrease selection bias, we preferred the tumour specimen collected from the primary site rather than the metastatic site. Newly obtained specimens were preferred to archived samples (90 days prior to start of PD-L1 IHC examination). In addition, a training programme on IHC analysis and CPS scoring was performed locally at each site by a trained and certified pathologist to reduce inter-group heterogeneity. Finally, imputation was performed to deal with the missing data bias related to the medical characteristic information.
According to a prevalence of 44% of HNSCC samples with PD-L1 (CPS ≥20) expression based on Chinese subgroup data from the KEYNOTE-048 study, a total sample size of 396 cases with HNSCC was required to estimate the PD-L1 expression prevalence with ±5% precision with a 95% CI.
The categorical variables are presented as count with proportion or percentage, and continuous variables are described as mean and SD. The prevalence of PD-L1 expression is presented in the form of percentage of patients with PD-L1 CPS ≥20 and with CPS ≥1, separately. In addition, 95% Clopper-Pearson intervals were calculated for the prevalence, to describe PD-L1 expression according to baseline demographic characteristics, clinicopathological characteristics, treatment status and other available biomarkers. The χ2 test and Cochran-Mantel-Haenszel χ2 test were used to compare the prevalence of related variables (eg, demographic characteristics, clinicopathological parameters and treatment status) among patients with different PD-L1 expression statuses to explore the potential factors associated with PD-L1 expression (CPS ≥20) in R/M HNSCC.
Study data are available from the corresponding author as appropriately required.
In the six participating sites, we screened a total of 406 patients with R/M HNSCC. No patient was excluded according to the inclusion/exclusion criteria. Of these patients, 123 had previously signed an ICF for use of her/his sample for future research and 283 were under conditions accepted by the IRB/ERC for waiving the ICF. During PD-L1 IHC analysis, four patients were excluded from the final analysis for insufficient viable tumour cells on the slides (figure 1).
Among all 402 patients, the median age at initial diagnosis was 56.8 years (range 26–91 years). The primary tumour site was the oral cavity for nearly half of the patients (189/402, 47.0%), and more than one-third (138/402, 34.3%) had an Eastern Cooperative Oncology Group Performance Status of 1. For most cases (380/402, 94.5%), samples were available from the primary tumour sites, while 14 samples were from metastatic lesion and 8 from lymph node. Among the patients, 27 had new biopsies while 375 had archival tissues within the last 2 years. The detailed baseline characteristics of the included patients are described in table 1.
Prevalence of PD-L1 expression
For all 402 examined patients with HNSCC, the prevalence of PD-L1 expression with a CPS ≥20 was 41.8% (95% CI: 36.9% to 46.8%), and the prevalence of PD-L1 expression with a CPS ≥1 was 83.8% (95% CI: 79.9% to 87.3%).
Potential associated factors
Between the PD-L1 CPS ≥20 and CPS <20 patient groups, statistically significant differences were observed in variables of sex (p<0.001), smoking habit (p=0.0138, non-smokers vs current smokers) and primary tumour site (p<0.001 for hypopharynx vs oral cavity and p=0.0304 for larynx vs oral cavity). On univariate logistic regression analyses, the prevalence of PD-L1 CPS ≥20 expression was significantly higher in women versus men (69.6% vs 37.3%; OR, 3.86; 95% CI: 2.097 to 7.101; p<0.001) and in non-smokers versus current smokers (50.8% vs 35.1%; OR, 1.91; 95% CI: 1.141 to 3.230; p=0.0138). Conversely, there was a lower prevalence of PD-L1 CPS ≥20 expression in patients whose primary tutor sites were the hypopharynx (18.6% vs 51.9%; OR, 0.21; 95% CI: 0.109 to 0.412; p<0.001) or larynx (38.4% vs 51.9%; OR, 0.58; 95% CI: 0.352 to 0.949; p=0.0304) compared with those whose primary tutor site was the oral cavity. And between the PD-L1 CPS ≥1 and CPS <1 patient groups, the percentage of patients with CPS ≥1 was higher among patients who had a history of surgery than among patients who did not have a history of surgery (86.2% vs 74.7%; OR, 2.12; 95% CI: 1.175 to 3.813; p=0.0125). The details of the comparisons between the PD-L1 CPS ≥20 and CPS <20 patient groups are listed in table 2 and online supplemental appendix 2.
The present EXCEED study revealed the prevalence of PD-L1 expression status in Chinese patients with R/M HNSCC, and we found that the prevalence of PD-L1 CPS ≥20 expression was 41.8% and that of PD-L1 CPS ≥1 expression was 83.8%, which was consistent with the findings of the KEYNOTE-055 study (82% for CPS ≥1%) and the KEYNOTE-048 study (43% for CPS ≥20 and 85% for CPS ≥1).18 22 The Food and Drug Administration(FDA) approved pembrolizumab (KEYTRUDA, Merck) for the first-line treatment of R/M HNSCC on 10 June 2019, for use in combination with platinum and fluorouracil (FU) for all patients and as a single agent for patients whose tumours express PD-L1 (CPS ≥1) as determined by an FDA-approved test, whereas the National Medicine Products Administration approved pembrolizumab monotherapy for treating advanced R/M HNSCC with PD-L1 CPS ≥20. Our findings filled in time of the data gap for PD-L1 expression in Chinese patients with R/M HNSCC and closely aligned with current indications in China, providing a solid foundation for healthcare professionals to develop suitable detection and treatment strategies and for future research. The present study further identified several potential factors associated with PD-L1 CPS ≥20 expression in R/M HNSCC, including sex, smoking habit and primary tumour site. The sex difference in PD-L1 expression was previously reported in patients with non-small cell lung cancer.28 However, no sex difference in PD-L1 expression in nasopharyngeal carcinoma was reported by Feng et al.29 Despite the inconsistent relationship between sex and PD-L1 expression, the sex-based differences in response to anti-PD-1 or PD-L1 treatment have been confirmed by systematic reviews.30 31 The rationale may be related to sex differences in innate and adaptive immune responses.32 Our findings revealed that the sex difference may be related to PD-L1 expression in Chinese patients with R/M HNSCC, which will be a good indicator for healthcare providers in making diagnosis and treatment decisions. Another factor associated with PD-L1 expression, smoking, has been reported to reduce PD-L1 expression in lung epithelial cells in vitro and in vivo.33 The rationale for the influence of smoking on PD-L1 expression may be related to the aryl hydrocarbon receptor (AhR), which can be affected by smoking and contributes to immunosuppression via regulation of important immune checkpoints (eg, AhR can mediate tobacco-induced PD-L1 expression and transcriptional control CD39 expression) in oral squamous cell carcinoma.34 We found that non-smokers had a higher PD-L1 (CPS ≥20) ratio than patients who currently smoke, which was consistent with the previous findings that PD-L1 expression appears to be higher in female patients, non-smokers, non-drinkers reported by Lenouvel et al.35 Finally, our findings indicate that the primary tumour sites of the hypopharynx and larynx may have less PD-L1 expression than the oral cavity, without existing evidence to support this statement, further research is needed to draw stronger conclusions.
The existing literature indicates that pathologists are currently facing several challenges when evaluating PD-L1 expression in patients with HNSCC. Among them, the most important one is concordance among assays.36 Despite a clear concomitant diagnosis (CDx) being approved, many laboratories work on other platforms and/or with other clones, which creates variability in detected PD-L1 expression in patients with HNSCC (ranges from 18% to 100%). The main published findings to date indicate that the degree of concordance varies from fair to substantial among the reference assay 22C3 pharmDx and two alternate Ventana assays (SP262 and SP142), while limited evidence is available concerning the 28-8 assay.36 Hence, further work is needed to draw stronger conclusions on the interchangeability of PD-L1 assays in HNSCC.36 37 In our study, we adopted 22C3 pharmDx on the Dako IHC platform. CPS was scored by trained and certified pathologists and provided reliable data concerning the prevalence of PD-L1 expression status in Chinese patients with R/M HNSCC.
The second challenge lies in concordance among pathologists. CPS’s apparent higher sensitivity at lower cutoffs of positivity supported its widespread adoption in testing patients with HNSCC. Indeed, CPS is more complex and less intuitive than Tumor Proportion Score (TPS), as it requires specific counting of tumour and immune cells to calculate the score.36 Not surprisingly, training could narrow both interobserver and intraobserver differences. Also, better concordance was found among trained pathologists, with the reproducibility being higher in assessing CPS.38 In addition, digital tools such as artificial intelligence and algorithms may serve certain important aid for pathologists evaluating PD-L1 expression, since automated algorithms can be trained to recognise different types of cells (tumour vs immune; with macrophages to be excluded from other cell types) and provide a more objective quantification of CPS.36 39
The third challenge is the intrinsic heterogeneity of PD-L1 expression in HNSCC, including temporal and spatial heterogeneity. First, PD-L1 expression is heterogeneous within primary tumours, thus choosing samples that are representative of the whole tumour is important. Concordance between PD-L1 expression in primary HNSCC and neck Lymph node metastasis (LNM) is fair to moderate, and PD-L1 may be over-expressed in LNM even in cases that have negative primary tumours. Consequently, clinicians should undertake a careful evaluation of PD-L1 both in primary HNSCC and on LNM if the latter is available.40 Furthermore, by comparing PD-L1 expression between freshly biopsied tissue and the same tissue later in time, Karpathiou et al provided evidence of PD-L1 expression diminishing, in both tumour and immune cells, over time in storage.41 Therefore, an immediate rather than a retrospective assay of PD-L1 expression should be preferable in routine practice. To minimise heterogeneity, we preferred that the tumour specimen be collected from the primary site rather than the metastatic site when possible, with 380 samples (94.5%) from the primary tumour, 8 samples (2.0%) from the lymph node and 14 samples (3.5%) from metastatic lesions, as relatively uniform sample sourcing could make the result more homogeneous. In addition, we aimed to consider the time factor by selecting newly obtained specimens preferred to archived samples (90 days prior to start of PD-L1 IHC examination). However, due to various limitations such as difficult cases and/or ethical considerations, only 27 specimens (6.7%) were newly obtained and the other 375 (93.3%) were archived. In an effort to compensate for the heterogeneity this may have caused, the most recent sample available was always selected whenever using archived samples, and none of them were older than 2 years.
The potential impact of already administered therapy on PD-L1 expression is also a concern.40 To date, findings regarding the effects of anti-tumour Chemotherapy (CT) and Radiation Therapy (RT) on PD-L1 expression in HNSCC have been conflicting.42–44 And history of systemic therapy and RT treatment hardly had an impact on PD-L1 expression in our study. PD-L1 expression in tumour cells of patients undergoing combined therapy with a platinum agent tends to be higher, though further studies are needed to clarify this issue.45–47 Moreover, in contrast to many previous studies using TPS which showed conflicting results, only a few recent papers and two USCAP abstracts used the CPS scoring method and judged the correlations between biopsy and resection specimens and between primary and metastasis sites as generally modest.40 48–50 However, Liu et al found that the concordance between FNA of metastatic lymph nodes and histological material was 76% when analysed with both CPS and TPS. In their study, the positive predictive value was 100% for both scoring assessments, but the negative predictive value was only 54% and 50%, with most of discordant cases being negative in histological material.51 The evaluation of PD-L1 expression using CPS has not been well established, mainly because of challenges in identifying tumour-associated immune cells in the absence of tissue architecture.37 Recent evidence suggests that many other markers and mechanisms are involved in the dysregulation of the cancer-related immune system and affect the expression of related markers such as PD-L1.52 It is then foreseeable that in the near future, growing evidence will provide more quantitative, reliable and practical evidence of PD-L1 expression in HNSCC.53 For the treatment history before bio-specimen collection, 319 (79.4%) of the included patients received surgical treatment, 51 (12.7%) received radiation treatment, 59 (14.7%) received chemotherapy, 11 (2.7%) received immunotherapy, 13 (3.2%) underwent targeted therapy and 1 patient (0.2%) was given other treatment. The results showed that patients who had a surgical treatment history (p=0.0125) had a higher PD-L1 (CPS ≥1) ratio than patients who had no surgical treatment history. Furthermore, we quite agree on the consistency of PD-L1 using CPS scoring between core biopsy and resection specimens, but not FNA cell block or cytological material, for PD-L1 positive immune cells, PD-L1 tends to gather at the tumour–stroma interface and be easily missed by FNA or cytological procedures.49 Therefore, the exclusion criteria in our study included that ‘availability of only a specimen obtained with fine needle aspirate (FNA) or cytologic specimen’, and more robust evidence is needed to draw clear conclusions.
It is now well demonstrated that alcohol consumption, smoking and viral infections, such as HPVs are major risk factors for HNSCC development. Particularly, HPV16 is responsible for more than 90% of virus-driven HNSCC, and the HPV-driven HNSCC incidence is increasing in Europe and North America.54–56 Findings in the literature about the association of HPV with PD-L1 expression and prognosis have yet to reach a consensus, and the present study explored the relationship between HPV infection and PD-L1 expression in HNSCC. Xu et al reported that of 112 HPV (+) oropharyngeal squamous cell carcinoma (OPSCC) tumours, high (CPS ≥20), intermediate (1≤CPS<20) and low (CPS <1) PD-L1 expression was seen in 29.5%, 43.8% and 26.8% Chinese cases, respectively.57 However, with limited data, PD-L1 expression was observed in the majority of patients with OPSCC regardless of HPV status. Moreover, a trend of higher PD-L1 (CPS ≥20) ratio in HPV-positive patients whose primary tumour site were oral cavity than HPV-negative ones was noted among the Chinese patients, and whether this is related to race or geography remains to be confirmed.
To our best knowledge, this study is the first clinical study aimed at exploring the PD-L1 expression condition in R/M HNSCC among Chinese patients. Our study provides solid insight for healthcare providers to use in making diagnosis and treatment decisions in Chinese R/M HNSCC cases. In addition, we applied the CPS to score PD-L1 expression, which can reflect the integrated condition under the specific R/M HSNSCC immune environment. However, this study still has some limitations. First, as this is a multicentre retrospective analysis, there may be variability in data quality and completeness across centres including inter-rater variability for PD-L1 scoring. Additionally, selection bias may be a potential concern. Also, even though proper training was provided to all participating pathologists, some amount of bias was inevitably present in the subjective process of imaging interpretation.
According to our findings, more than 40% of Chinese patients with R/M HNSCC had PD-L1 expression with a CPS ≥20, which was significantly associated with sex, smoking history, and primary tumour site. Our findings regarding the variables related to the PD-L1 expression level provide insight for clinical practice and a solid basis for future research on the use of immunotherapy in HNSCC.
Data availability statement
Data are available upon reasonable request. The data generated in this study are available upon request from the corresponding author.
Patient consent for publication
This study involves human participants and was approved. The protocol and protocol amendment of this study were reviewed and approved by each study site’s Independent Ethics Committee (Ethics Committee of National Cancer Center/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College (approval no. 20/377-2573), Ethics Committee of Tongji Hospital,Tongji Medical College, Huazhong University of Science & Technology (approval no. 2020-S332), Ethics Committee of Biomedical Research, West China Hospital of Sichuan University (approval no. 2021-343), Ethics Committee of Sun Yat-Sen University Cancer Center (approval no. B2021-087-01), Ethics Committee of Henan Cancer Hospital (approval no. 2020103012-001), Ethics Committee of Fudan University Shanghai Cancer Center (approval no. 2102231-16)). Participants gave informed consent to participate in the study before taking part.
Editorial assistance was provided by Medjaden, Inc. This assistance was funded by MSD China.
Handling editor Munita Bal.
HL, DK, PZ and JZ contributed equally.
Contributors JY is responsible for the overall content as the guarantor. HL, JW, JL, WT, JC and JY contributed to conception, design or planning of the study. HL, DK, PZ, JZ, QX, PD, LJ, SZ, YJ, XJ and JY contributed to acquisition of the data. HL, YJ, XJ, JZ and JY contributed to analysis of the data. HL, LJ, JZ, JC and JY contributed to interpretation of the results. HL and JY contributed to drafting of the manuscript. All authors contributed to reviewing or revising the manuscript for important intellectual content. HL, LJ, SZ, YJ and XJ contributed to provision of study materials. JW and WT contributed to administrative, logistical or technical support. All authors read and approved the final version of the manuscript.
Funding The study was funded by MSD China.
Competing interests JL, WT, JZ and JC are employees of MSD China; all other authors have none to declare.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.